Earlier today NVIDIA Corporation launched its two new graphics processors for high-end and performance mainstream markets, the GeForce FX 5950 Ultra and the GeForce FX 5700 Ultra. The former is just an overclocked GeForce FX 5900 GPU, while the latter is NVIDIA’s first commercial graphics processor made at IBM’s East Fishkill, New York, foundry using 0.13 micron fabrication process.

NVIDIA GeForce FX 5700 and 5700 Ultra has 4 rendering pipelines, just like the predecessor, but also boasts with improved pixel and vertex shader processing engines, dubbed CineFX 2.0, as well as NVIDIA UltraShadow feature. It functions at blazing 475MHz, as high as the new high-end monster, the GeForce FX 5950 Ultra.

In order to find out more about IBM’s first graphics processing unit for NVIDIA, we got eVGA’s e-GeForce FX 5700 Ultra Limited Edition graphics card featuring the milestone GPU from the Santa Clara, California-based firm.

The graphics card itself looks and works great, it features 900MHz GDDR2 128-bit memory with peak bandwidth of up to 14.40GB/s. The PCB of the GeForce FX 5700 Ultra is as huge as the PCB for the GeForce FX 5900, even though the design of the former is not as complicated as of that of the 5900.

What we were really curious about was the chip itself. We could not wait to see the IBM logotype on the core that is mounted on the FCBGA substrate and thus can be seen easily, just like all ATI’s cores since the legendary RADEON 9700 PRO.

To our disappointment, we saw a core with “Korea” mark and no IBM logotypes on it. In fact, we already mentioned this year that NVIDIA either packages or manufactures the GeForce FX 5600 Ultra in Korea, as it also had the same marking. Now we can conclude that the new GeForce FX 5700 Ultra also has something to do with Korean facilities, as its predecessor. Though, we cannot tell you what exactly NVIDIA does with its chips in the South Korea, it is a bit surpising to see mention about Korea on a chip made in the USA…